Re: Controlling water diffusion

From: Peter Freddolino (petefred_at_umich.edu)
Date: Sun Jul 19 2015 - 21:13:15 CDT

Dear Evandro,

Let’s be clear, this behavior should not be different with the charmm or amber force field. If you have properly configured periodic boundary conditions, this is just an effect of how you’ve set up coordinate wrapping.

Also, PME (particle mesh Ewald) requires PBCs (periodic boundary conditions), but the inverse is not true. PME just has to do with long range electrostatics calculations, and will not alter wrapping. Please do not confuse these two things.

You need to carefully inspect your system to make sure that when properly viewing it in the presence of the periodic boundaries you have configured, it is correct; I can’t make any guarantees just on the basis of hearing about a case, but I can tell you that the symptoms you’ve described are usually just because people don’t realize the effects of various choices of wrapping parameters.

Best,
Peter

> On Jul 17, 2015, at 8:44 AM, Evandro Semighini <epsemighini_at_gmail.com> wrote:
>
> Hello Peter,
>
> Thank you !
>
> I really thought this strange, because using CHARMM FF this two things never happened to me or my co-workers, and because PME was always set in my simulations, so, I was expecting that the part of the protein that gone out of the box would be "teleported" to the other side of the system. I always imagined how it will mess up RMSD measurements.
>
> Yes, I saw the deformations on the water box you said, and I imagined it happened due to the empty space left by the protein at that volume.
>
> So, it's more a issue with visualization of the system, since I am using PME, and the protein goes trough the MD normally?
> Thank you for the tip of pbctools and periodic images, I saw it yesterday was waiting the "out of box" run finish to have more material to work with.
>
> Thank you again !
> Evandro
>
> 2015-07-16 23:54 GMT-03:00 Peter Freddolino <petefred_at_umich.edu>:
> Hi Evandro,
> The key here is that you’re simulating a periodic system. You don’t have just one box, but an infinite number of identical boxes tiled through space. So when your waters appear to go out of the box when wrap is off, they’ve just gone into the next periodic copy over (but it is replaced by a copy coming in through the other side, so the contents of your unit cell are constant). When you have wrap on, your protein may appear to stick out of the box because it hasn’t moved far enough to be wrapped. You should see a complementary gap in the water on the opposite side of the system where it is occupying space. Viewing your system with the help of the pbctools plugin, or with periodic images turned on (in the ‘periodic’ tab of the graphical representations menu), can help to see what the system really looks like.
> Best,
> Peter
>
> > On Jul 16, 2015, at 3:21 PM, Evandro Semighini <epsemighini_at_gmail.com> wrote:
> >
> > Hello,
> >
> > I am running NAMD using AMBER Force Field and I am facing a issue with the water box.
> >
> > I measured the water vectors and size with VMD and configured the periodic boundary conditions as described at the tutorials, and ran a simulation, without defining PMEGridSize and using wrap comands.
> > The simulation goes well, except that the water molecules diffuses a lot away from the box, they seems to evaporate.
> >
> > I used then the wrapWater and wrapAll comands and things got more strange, as the protein is entirely removed from the water box after energy minimization, but the MD runs ok after that.
> >
> > I checked the data from the cell vectors and center, as well as the coordinate files and the numbers are within the expected values.
> >
> > The minimization seems to reset the coordinates of the protein.
> >
> > Can anyone, please, point me where I am missing ?
> >
> > Thank you,
> > Evandro
>
>

This archive was generated by hypermail 2.1.6 : Thu Dec 31 2015 - 23:21:59 CST